The Journal of Experimental Biology 208, 3015-3035 3015 Published by The Company of Biologists 2005 doi:10.1242/jeb.01745 Commentary Phylogenetic approaches in comparative physiology Theodore Garland, Jr1, Albert F. Bennett2,* and Enrico L. Rezende1 1Department of Biology, University of California, Riverside, CA 92521, USA and 2Department of Ecology and Evolutionary Biology, University of California, Irvine, CA 92697, USA *Author for correspondence (e-mail: [email protected]) Accepted 13 June 2005 Summary Over the past two decades, comparative biological and Monte Carlo computer simulations). We discuss when analyses have undergone profound changes with the and how to use phylogenetic information in comparative incorporation of rigorous evolutionary perspectives and studies and provide several examples in which it has been phylogenetic information. This change followed in large helpful, or even crucial, to a comparative analysis. We also part from the realization that traditional methods of consider some difficulties with phylogenetically based statistical analysis tacitly assumed independence of all statistical methods, and of comparative approaches in observations, when in fact biological groups such as general, both practical and theoretical. It is our personal species are differentially related to each other according to opinion that the incorporation of phylogeny information their evolutionary history. New phylogenetically based into comparative studies has been highly beneficial, not analytical methods were then rapidly developed, only because it can improve the reliability of statistical incorporated into ‘the comparative method’, and applied inferences, but also because it continually emphasizes the to many physiological, biochemical, morphological and potential importance of past evolutionary history in behavioral investigations. We now review the rationale for determining current form and function. including phylogenetic information in comparative studies and briefly discuss three methods for doing this Key words: allometry, comparative method, evolutionary physiology, (independent contrasts, generalized least-squares models, model of evolution, phylogeny, statistical analysis. Introduction Studies of organismal form and function rely on multiple evolution across clades or the amount of morphospace types of scientific investigation, including theory, description, occupied by clades or by ecologically defined groups (Garland, experimentation and comparison. Comparing species is an 1992; Clobert et al., 1998; Ricklefs and Nealen, 1998; Garland ancient human enterprise, done for a variety of reasons and Ives, 2000; Hutcheon and Garland, 2004; McKechnie and (Sanford et al., 2002). Since Charles Darwin, the ‘comparative Wolf, 2004). Of particular interest for the present review, they method’ – comparing populations, species or higher taxa – has are also widely used to explore trade-offs (e.g. Clobert et al., been the most common and productive means of elucidating 1998; Vanhooydonck and Van Damme, 2001) and to examine past evolutionary processes (Harvey and Pagel, 1991; Brooks functional (mechanistic) relationships among traits (e.g. and McClennan, 2002). Comparative methods have been used Lauder, 1990; Iwaniuk et al., 1999; Mottishaw et al., 1999; extensively to infer evolutionary adaptation, that is, changes in Autumn et al., 2002; Hale et al., 2002; Gibbs et al., 2003; response to natural selection (for alternate physiological Johnston et al., 2003; Herrel et al., 2005), including allometric meanings of ‘adaptation’, see Garland and Adolph, 1991; scaling with body size (e.g. Garland, 1994; Reynolds and Lee, Bennett, 1997). They are most often promoted and criticized 1996; Williams, 1996; Clobert et al., 1998; Garland and Ives, (e.g. Leroi et al., 1994) within this context. However, 2000; Nunn and Barton, 2000; Herrel et al., 2002; Perry and comparative methods are not used to infer adaptation alone Garland, 2002; Rezende et al., 2002, 2004; Schleucher and (Garland and Adolph, 1994; Sanford et al., 2002), but are also Withers, 2002; McGuire, 2003; Al-kahtani et al., 2004; employed to analyze the effects of sexual selection (e.g. McKechnie and Wolf, 2004; Muñoz-Garcia and Williams, in Hosken et al., 2001; Nunn, 2002; Smith and Cheverud, 2002; press). Aparicio et al., 2003; Cox et al., 2003), which may be Comparative methods have been radically restructured over nonadaptive or even maladaptive with respect to natural the past two decades, and now routinely incorporate both selection. These methods can also be used to compare rates of phylogenetic information and explicit models of character THE JOURNAL OF EXPERIMENTAL BIOLOGY 3016 T. Garland, A. F. Bennett and E. L. Rezende evolution. Indeed, Sanford et al. (2002) suggest that this new Maddison, 2000), and in Mesquite (http://mesquiteproject.org/ emphasis be termed the ‘comparative phylogenetic method’. mesquite/mesquite.html; see also Paradis and Claude, 2002). As outlined in Blomberg and Garland (2002), this revolution For a general listing of phylogeny-related programs, see in comparative phylogenetic methodology followed from the website maintained by Joe Felsenstein (http:// several conceptual advances: (1) adaptation should not be evolution.genetics.washington.edu/phylip/software.html). casually inferred from comparative data; (2) the incorporation We discuss when phylogenetically based statistical methods of phylogenetic information increases both the quality and should be used and give some practical examples of where a even the type of inference from comparative data alone; (3) phylogenetic perspective has improved our understanding of because all organisms are differentially related to each other, comparative data and evolutionary processes. We also discuss taxa cannot be assumed to be independent of each other for some of the practical and theoretical limitations of such statistical purposes; (4) statistical analyses of comparative data methods. Throughout, we try to emphasize that the must assume some model of character evolution for effective incorporation of phylogeny can greatly enhance comparative inference; (5) taxa used in comparative analyses should be studies, deliver new insights, and open new areas for research. chosen in regard to their phylogenetic affinities as well as the This is of necessity only a brief summary and readers are area of functional investigation; and (6) even phylogenetically directed to more extensive discussions of the topics and issues based comparisons are purely correlational and inferences of raised here (e.g. Ridley, 1983; Lauder, 1981, 1982, 1990; causation drawn from them can be enhanced by other Harvey and Pagel, 1991; Garland et al., 1992, 1999; Garland approaches, including experimental manipulations. and Adolph, 1994; Harvey, 1996; Ricklefs and Nealen, 1998; To expand on some of these points, ‘quality’ in point 2 Ackerly, 1999, 2000, 2004; Pagel, 1999; Purvis and Webster, includes the simple fact that adding an independent estimate of 1999; Diniz-Filho, 2000; Feder et al., 2000; Garland and Ives, phylogenetic relationships to a comparative analysis increases 2000; Maddison and Maddison, 2000; Garland, 2001; Rohlf, – often greatly – the amount of basic data that is brought to 2001; Autumn et al., 2002; Blomberg and Garland, 2002; bear on a given question, whereas ‘type’ refers to analyses that Brooks and McLennan, 2002; Blomberg et al., 2002, 2003; are simply impossible without a phylogenetic perspective, such Rezende and Garland, 2003; Housworth et al., 2004). We have as reconstructing ancestral values or comparing rates of intentionally not cited some ‘forum’ and ‘perspective’ type evolution among lineages. Although phylogenetic information papers because we felt that their rhetoric was misleading, and and a suitable analytical method may allow any comparative in some cases they contain outright errors. data set to be ‘rescued’ from phylogenetic nonindependence The empirical examples cited here are idiosyncratic, (e.g. avoid inflated Type I error rates; point 3), reflecting mainly our own research interests. Thus, we phylogenetically informed choice of species (point 5) can emphasize examples that involve physiological phenotypes, accomplish more, such as actually increasing statistical power but include others when they are lacking. Our enthusiasm for to detect relationships among traits (Garland et al., 1993; phylogenetic approaches in comparative physiology should not Garland, 2001). Finally, we note that point 6 was recognized be taken to imply, however, that we think they are more long ago, but has been re-emphasized as phylogenetically important than other approaches, such as measurement of explicit methods of statistical inference have been developed selection acting in natural populations, experimental evolution (e.g. see Lauder, 1990; Garland and Adolph, 1994; Leroi et al., (e.g. see Garland and Carter, 1994; Bennett and Lenski, 1999; 1994; Autumn et al., 2002). Ackerly et al., 2000; Feder et al., 2000; Garland, 2001, 2003; The intent of this commentary is to provide a review of some Bennett, 2003; Swallow and Garland, 2005), or more purely advances that have occurred in the comparative method, with mechanistic investigations (e.g. Mangum and Hochachka, an emphasis on their place in comparative physiology. We 1998; Hochachka and Somero, 2002). examine the underlying reasons for the incorporation of We are concerned that some of our discussion of phylogenetic information into comparative studies. In an assumptions